Tuberculosis, caused by infection with Mycobacterium tuberculosis, remains a serious threat to global health. Mycobacterium tuberculosis latently infects a third of the world and kills millions every year. Treating tuberculosis remains difficult. The drug regimen involves four antibiotics and a minimum of six months of treatment. Though antibiotic treatment kills a significant portion of the bacterial population quickly, some cells are able to tolerate treatment for extended periods of time and therefore necessitate the long drug exposure. The arduous treatment makes compliance and cure difficult, leading to transmission and emergence of drug tolerant strains. We therefore have a dire need to design shorter, more effective treatments against tuberculosis. The basic treatment for tuberculosis has not been significantly improved in decades. This failure is due, in part, to a lack of understanding of the features of drug tolerant subpopulations. Here, we seek to overcome this obstacle by characterizing these important subpopulations and using quantitative descriptions of their physiology and response to antibiotic treatment to rationally design improved drug regimens. We have recently developed a microfluidics-based live cell microscopy system to study the growth properties and antibiotic response of individual mycobacteria. Using our system, we discovered that an unusual pattern of unipolar growth creates variability in the growth rate and antibiotic susceptibility of mycobacteria. We propose that this pattern of asymmetric growth is a broad mechanism that creates subpopulations of cells with distinct physiological properties that make them differentially tolerant to specific classes of antibiotics and host induced stress. Here, we will combine our live-cell imaging system with automatic image analysis and mathematical modeling to determine the relationship between antibiotic response and the cell growth and cell cycle characteristics of individual Mycobacterium tuberculosis cells. We will merge linear regression models with novel semi-mechanistic pharmacodynamic models that translate single cell descriptions into population behavior. We will use our models and experimental system to design and test new dosing schedules that effectively target the generation and survival of drug- and immune-tolerant mycobacterial subpopulations. We anticipate that this study will provide a foundational framework to systematically engineer improved clinical tuberculosis treatments while also establishing a broadly applicable methodology to rationally design improved therapies for other diseases.